Re: Are there "compiler generators"? (Anton Ertl)
Sun, 29 May 2022 06:45:21 GMT

          From comp.compilers

Related articles
Are there "compiler generators"? (Roger L Costello) (2022-05-28)
Re: Are there "compiler generators"? (Robin Vowels) (2022-05-29)
Re: Are there "compiler generators"? (Jan Ziak) (2022-05-28)
Re: Are there "compiler generators"? (2022-05-29)
Re: Are there "compiler generators"? (Thomas Koenig) (2022-05-29)
Re: Are there "compiler generators"? (Martin Ward) (2022-05-29)
Re: Are there "compiler generators"? (Fernando) (2022-05-29)
Re: Are there "compiler generators"? (gah4) (2022-05-29)
Re: Are there "compiler generators"? (2022-05-30)
Re: Are there "compiler generators"? (Hans-Peter Diettrich) (2022-05-30)
[9 later articles]
| List of all articles for this month |

From: (Anton Ertl)
Newsgroups: comp.compilers
Date: Sun, 29 May 2022 06:45:21 GMT
Organization: Institut fuer Computersprachen, Technische Universitaet Wien
References: 22-05-054
Injection-Info:; posting-host=""; logging-data="53684"; mail-complaints-to=""
Keywords: tools
Posted-Date: 29 May 2022 18:14:55 EDT

Roger L Costello <> writes:
>Are there compiler generators?

Apart from scanner generators and parser generators, there are also:

* generators of attribute grammar evaluators (e.g., Ox). They use an
    attribute grammar as input.

* generators of instruction selectors (e.g., burg and friends). Burg
    uses a tree grammar as input.

* generators of virtual machine interpreters and other code dealing
    with virtual machine instructions (e.g., vmgen). Vmgen uses a
    description of the virtual machine instructions as inputs.

* There is also some work on generating type checkers.

All of these could be classified as non-procedural, although they tend
to include some procedural code.

You need to add substantial amounts of glue code to integrate the
pieces generated by these generators into a compiler, especially for
languages that do not quite fit in the generators' molds. Plus, there
are pieces such as symbol tables and register allocation that are not
covered by these generators.

I have also seen a paper by a French group (don't remember the names)
in the early 1990s where a generator could generate a complete Pascal
compiler from a specification. My impression, however, was that the
generator could only generate compilers for languages that are
relatively close to Pascal, and I saw no good way to make it much more

- anton
M. Anton Ertl

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.